Create a new Kedro project¶

There are a few ways to create a new project once you have installed Kedro. For example, you can create a basic Kedro project set up with the project directories and basic code, but empty to extend as you need. Alternatively, you can create a Kedro project populated with template code that acts as a starter example.

Create a new empty project¶

The simplest way to create a default Kedro project is to navigate to your preferred directory and type:

kedro new

You will be asked to enter a name for your project, which can be human-readable and may contain alphanumeric symbols, spaces, underscores and hyphens. It must be at least two characters long.

Your choice is set as the value of project_name and is used to generate the repo_name and python_package automatically.

So, if you enter “Get Started”, the directory name for the project (repo_name) is automatically set to be get-started, and the Python package name (python_package) for your project is set to be get_started.

Description	Setting	Example
A human-readable name for your new project	`project_name`	`Get Started`
Local directory to store your project	`repo_name`	`get-started`
The Python package name for your project (short, all-lowercase)	`python_package`	`get_started`

The output of kedro new is a directory containing all the project files and subdirectories required for a basic Kedro project, ready to extend with your own code.

Create a new project from a configuration file¶

If you prefer to customise your new project’s directory and package name, you can instead use a configuration file to specify those values. The configuration file must contain:

output_dir The path in which to create the project directory
project_name
repo_name
python_package

The output_dir can be set to wherever you want to create the project. For example, ~ for your home directory, or . for the current working directory. Here is an example config.yml, which assumes that a directory named ~/code already exists:

output_dir: ~/code
project_name: My First Kedro Project
repo_name: testing-kedro
python_package: test_kedro

To create this new project:

kedro new --config <path>/config.yml

Create a new project containing example code¶

You can use a Kedro Starter to create a project containing template code, to run as-is or to adapt and extend.

To illustrate, we will create a Kedro project with example code based on the familiar Iris dataset.

Background information for the iris dataset example¶

The dataset was generated in 1936 by the British statistician and biologist Ronald Fisher. The dataset contains 150 samples in total, comprising 50 samples of 3 different species of Iris plant (Iris Setosa, Iris Versicolour and Iris Virginica). For each sample, the flower measurements are recorded for the sepal length, sepal width, petal length and petal width.

A machine learning model can use the Iris dataset to illustrate classification (a method used to determine the type of an object by comparison with similar objects that have previously been categorised). Once trained on known data, the machine learning model can make a predictive classification by comparing a test object to the output of its training data.

Create the example project¶

The first step is to create the Kedro project using a starter to add the example code and data. Feel free to name your project as you like, but here we will assume the project’s name is get started.

kedro new --starter=pandas-iris

Run the example project¶

Once you have created the project, to run project-specific Kedro commands, you must navigate to the directory in which it has been created and install the project’s dependencies:

cd get-started
pip install -r src/requirements.txt

You are ready to run the project:

kedro run

Note

The first time you type a kedro command in your new project, you will be asked whether you wish to opt into usage analytics. Your decision is recorded in the .telemetry file so that subsequent calls to kedro in this project do not ask you again.

When the command completes, you should see a log message similar to the following in your console:

[08/09/22 11:23:30] INFO     Model has accuracy of 0.933 on test data.                                        nodes.py:74
                    INFO     Saving data to 'metrics' (MetricsDataSet)...                             data_catalog.py:382
                    INFO     Completed 3 out of 3 tasks                                           sequential_runner.py:85
                    INFO     Pipeline execution completed successfully.                                      runner.py:89

Under the hood: Pipelines and nodes¶

The example project contains a single pipeline stored in src/get_started/pipeline.py. The pipeline is comprised of nodes that are responsible for splitting the data into training and testing samples, running the 1-nearest neighbour algorithm to make predictions and accuracy-reporting.

The nodes are stored in src/get_started/nodes.py:

Node	Description	Node function name
Split data	Splits the example Iris dataset into train and test samples	`split_data`
Make Predictions	Makes class predictions (using 1-nearest neighbour classifier and train-test set)	`make_predictions`
Report accuracy	Reports the accuracy of the predictions performed by the previous node.	`report_accuracy`

Visualise the project¶

This is a swift introduction to show how to visualise the project with Kedro-Viz. See the visualisation documentation for more detail.

In your terminal type the following:

kedro viz

This command automatically opens a browser tab to serve the visualisation at http://127.0.0.1:4141/.

You should see the following, which you can explore to learn more about the pipeline, nodes and datasets:

To exit the visualisation, close the browser tab. To regain control of the terminal, enter ⌘+c on Mac or Ctrl+c on Windows or Linux machines.